How to lock a CPU's voltage PERMANTENTLY or at least put a cap on it?

[mrbeanladen]

I set my voltage at 1.17, it's perfectly stable in BIOS on an i5 2400 running at 3.8/3.7/3.7/3.6. But the idiotic windows 7 keeps changing it apparently, in BIOS it shows up exactly as I set it at but in windows it fluctuates. I thought it was because of EIST And all of that but it isn't, disabled all of that thermal management sheit but it still fluctuates like mad between 1.2-1.35 and 1.35 from what I've heard is good for an i7 2600k running at 4ghz+, way overkill for what I do... and in bios it's perfectly stable almost 0.2v lower so why add extra heat? How the hell can I fix this?
I also have another problem BTW, my FSB fluctuates between 100.29 and 100.32, it should be at default which is 100. If I turn bclk modifying on and change it to 100 i occasionally experience reboots during POST and bios resets itself. WTF? is this normal? I heard sandy bridge doesn't like you tampering with the bclk too much...

 

[Guest]

are you having a problem with temps?
if not, why the anal retentiveness? no offense i understand the desire to tweak things but it is relevant that "if it ain't broke don't fix it."

and it would be most helpful to know what motherboard you have.

 

[mrbeanladen]

are you having a problem with temps?
if not, why the anal retentiveness? no offense i understand the desire to tweak things but it is relevant that "if it ain't broke don't fix it."

and it would be most helpful to know what motherboard you have.

oh sorry i forgot it's a gigabyte P67a-D3-B3 with F7 BIOS
well, yeah temps are pretty high I guess, 70 degrees celsius in prime95 at 3.6ghz and on a new cooler (stock was 80-90) and there's no reason it should run that high, it doesn't need so much voltage it can manage with 0.2 less so why add more? don't want to lower my cpu's lifespan and cause heat for nothing.

 

[Guest]

70c is very good!
i have lowered my off set (or raised it with a negative value actually) for my Vcore in the bios. my voltage will fluctuate between .081 to 1.09 for my i3 2120. the thermal management actually helps by lowering the multi and volts when you idle; so turning it off is defeating your purpose.

all Vcore reading are not 100% accurate. i lowered the Vcore by raising the negative offset until i couldn't post; gave it a bit more juice and ran prime95 to see if it would fail (it did then gave just a little more juice and passed). what my bios reading is and coretemp and HW monitor reads all are different but i am confident i have the lowest stable voltage.

disabling speedstep and all the other thermal management is for people who want to overclock and kep the voltage more constant to increase stability. under volting is a bit of the opposite.

so i suggest you enable the thermal management, lower the voltage by raising the negative value (-) for the offset. boot, test in prime95 and see what that gets you. btw a pencil and paper is handy to keep track of what the value is in case you don't post and have to reset the bios.

hope this helped w/o getting long winded. but just don't worry about having the same exactly voltage across the board. it will fluctuate to stay stable because the load on it will fluctuate.

 

[Guest]

oh sorry i forgot it's a gigabyte P67a-D3-B3 with F7 BIOS
well, yeah temps are pretty high I guess, 70 degrees celsius in prime95 at 3.6ghz and on a new cooler (stock was 80-90) and there's no reason it should run that high, it doesn't need so much voltage it can manage with 0.2 less so why add more? don't want to lower my cpu's lifespan and cause heat for nothing.

Yes if you are in stock is fine one screw up in thermal paste and got 89 c...

 

[Guest]

OPS what is the new cooler ? 70 is high for a new cooler.

Ps Anyway the frequencies in cpu and all over the motherboard is created by crystals (piezoelectric). When electric voltage is pas to them they are start to vibrating producing a sine wave and the clock signal is moved to all the part by the cpu. The voltage is not always stable this why you see changes because it cannot be stable.

 

[mrbeanladen]

OPS what is the new cooler ? 70 is high for a new cooler.

Ps Anyway the frequencies in cpu and all over the motherboard is created by crystals (piezoelectric). When electric voltage is pas to them they are start to vibrating producing a sine wave and the clock signal is moved to all the part by the cpu. The voltage is not always stable this why you see changes because it cannot be stable.

I got this: http://www.coolermaster.co.uk/product.php?product_id=6776
and used arctic silver mx-2, put a very tiny pea in the middle very small just to get a nice circle in the middle and pressed hard on the heatsink and screwed it in strongly. I think it's the voltage's fault here, BIOS tells me that it runs at 1.17-1.18 and it does so perfectly but in windows the OS seems to don't give a shyte about voltage and it fluctuates between 1.2-1.35. Without turbo on it's about 1.07 or something and is about 10 degrees colder at load in prime95. I disabled those thermal... stuff and I noticed no change except that without them the clock is locked and sometimes it doesn't even stay that way... Should I just overclock from the main multiplier? maybe turbo sucks.
Also my room is about 30 celsius.

 

[Guest]

You can't do nothing about the voltage nobady has stable, in me the cpu is playing from 3006.18 to 3005.92, i get cpu up to 3.3 Gh with turbo boost on 100% load to the cores (it helps). In there the voltage is always more stable becaouse the cpu is draging stronger voltage current.

If you OC with the multiplier you are goanna get higher frequency for sure and more stable but is always nice to try to use and the turbo boost if it’s possible. If you max out the multiplier is nice to close the turbo boost in order to get stability to the system because if you don’t you must increase the voltage above the lowest operational you can find add get more heat to the cores. A 19% from turbo boost isn’t worth anyway.

For the heat is responsible your cpu cooler is blowing the heat to the motherboard and not to the rear out of the box. The heat is spreads faster to all the components on board before finally gets out.

A cpu cooler that blows out is always better.

 

[mrbeanladen]

You can't do nothing about the voltage nobady has stable, in me the cpu is playing from 3006.18 to 3005.92, i get cpu up to 3.3 Gh with turbo boost on 100% load to the cores (it helps). In there the voltage is always more stable becaouse becouse the cpu is draging stronger voltage current.

If you OC with the multiplier you are goanna get higher frequency for sure and more stable but is always nice to try to use and the turbo boost if it’s possible. If you max out the multiplier is nice to close the turbo boost in order to get stability to the system because if you don’t you must increase the voltage above the lowest operational you can find add get more heat to the cores. A 19% from turbo boost isn’t worth anyway.

For the heat is responsible your cpu cooler is blowing the heat to the motherboard and not to the rear out of the box. The heat is spreads faster to all the components on board before finally gets out.

A cpu cooler that blows out is always better.

Thanks.
Well, actually turbo has the max OC possible with my CPU, and even has a higher frequency at lower core number usage for example with 1 used core I get 3.8, with 2 3.7. When I set 3.7 in the multiplier I only get the max that turbo has too for 4 cores, which is 3.6. I did try it a few days ago it seems that It fixed my problem, it stayed at 1.23 volts and the temps were much lower almost 10 degrees lower and it has the same performance, 98-99 GFlops. I'll test it more tomorrow.

 

[mrbeanladen]

Best answer selected by mrbeanladen.